<p> In this thesis a non-parametric normalized stochastic approximation algorithm has been developed for the identification of multivariable systems from noisy data without prior knowledge of the statistics of measurement noise.</p> <p> The system model is first transformed into a special canonical form, then it is formulated in a non-parametric form. The parameters of this model are estimated through a normalized stochastic approximation algorithm. Finally, the system parameters are recovered from these estimates by another transformation.</p> <p> The proposed algorithm is applied to the identification of two simulated systems.</p> <p> Conclusions of this work and suggestions for future work are given.</p> / Thesis / Master of Engineering (MEngr)

The selection of a suitable model for efficiency analysis is one of the most important issues in policy analysis. Given the recent interest in the use of distance functions as alternative representation of production technology, this study compares the empirical performances of the parametric stochastic input distance function to its nonparametric counterpart, data envelopment analysis. A further comparison is made between the alternatives of a distance and production function frontiers. It further integrates efficiency scores from the consistent approaches in order to evaluate the performance of the sampled farm households and for analysis of policy impacts on technical, allocative and cost efficiency. The usefulness of the proposed methodology is applied to smallholder maize production in Benue State Nigeria. The maize subsector has featured in a number of Nigeria’s policy initiatives, the most current of which involves doubling of its production and productivity through promotion of improved technologies such as hybrid seed, inorganic fertilizer, pesticides, herbicides, and better management practices. Despite the policy initiatives, maize productivity has remained low raising questions about the efficiency of resource use by farmers and the benefits of Nigeria’s technology policy. The study used data obtained from a field survey for the 2008/2009 agricultural year. A multistage stratified sampling technique was employed in selection of respondents. A total of 240 maize farm households were randomly selected and interviewed using structured questionnaires. Results from all the approaches indicated considerable technical, allocative and cost inefficiency under both traditional and improved maize technology. Technical efficiency estimates range from 80 to 87 percent. Allocative efficiency estimates range from 53 to 74 percent while cost efficiency estimates range from 45 to 62 percent. The results from all the approaches indicated that inefficiency in maize production in Benue State is dominated by cost inefficiency suggesting the immense potential of enhancing production through improvement in overall efficiency. The overall consistency check shows that technical, allocative and cost efficiency measures from the three distance functions were consistent whereas similar conclusions could not hold when these were compared to the production frontier especially for technical efficiency estimates. Given the consistency of results from the parametric and non-parametric distance functions, an integrated input distance model was developed for providing final efficiency estimates and analysis of policy impacts. The results show that both traditional and improved technology users were technically, allocatively and cost inefficient. The average technical, allocative and cost efficiency are 84.2, 65.7 and 54.5 percent, respectively implying that there is a possibility of raising maize production by 45.5 percent through overall efficiency improvement. Under the integrated approach, the study revealed that hybrid seeds, inorganic fertilizer and conservation practices have positive and significant impact on farm efficiency. Other determinants of efficiency include education, age, household size, land size, credit, and membership in a farmer group. The findings justify the need for further public investment in maize technology development and proper implementation of the relevant policies in order to enhance the efficiency with which maize has been produced thereby increasing its productivity, food security and farm incomes and subsequently reducing poverty in Nigeria. / Thesis (PhD)--University of Pretoria, 2011. / Agricultural Economics, Extension and Rural Development / unrestricted